ATF7IP2/MCAF2 directs H3K9 methylation and meiotic gene regulation in the male germline.

H3K9 trimethylation (H3K9me3) plays emerging roles in gene regulation, beyond its accumulation on pericentric constitutive heterochromatin. It remains a mystery why and how H3K9me3 undergoes dynamic regulation in male meiosis. Here, we identify a novel, critical regulator of H3K9 methylation and spermatogenic heterochromatin organization: the germline-specific protein ATF7IP2 (MCAF2). We show that in male meiosis, ATF7IP2 amasses on autosomal and X-pericentric heterochromatin, spreads through the entirety of the sex chromosomes, and accumulates on thousands of autosomal promoters and retrotransposon loci. On the sex chromosomes, which undergo meiotic sex chromosome inactivation (MSCI), the DNA damage response pathway recruits ATF7IP2 to X-pericentric heterochromatin, where it facilitates the recruitment of SETDB1, a histone methyltransferase that catalyzes H3K9me3. In the absence of ATF7IP2, male germ cells are arrested in meiotic prophase I. Analyses of ATF7IP2-deficient meiosis reveal the protein's essential roles in the maintenance of MSCI, suppression of retrotransposons, and global up-regulation of autosomal genes. We propose that ATF7IP2 is a downstream effector of the DDR pathway in meiosis that coordinates the organization of heterochromatin and gene regulation through the spatial regulation of SETDB1-mediated H3K9me3 deposition.

Pregnancy enables antibody protection against intracellular infection.

Adaptive immune components are thought to exert non-overlapping roles in antimicrobial host defence, with antibodies targeting pathogens in the extracellular environment and T cells eliminating infection inside cells1,2. Reliance on antibodies for vertically transferred immunity from mothers to babies may explain neonatal susceptibility to intracellular infections3,4. Here we show that pregnancy-induced post-translational antibody modification enables protection against the prototypical intracellular pathogen Listeria monocytogenes. Infection susceptibility was reversed in neonatal mice born to preconceptually primed mothers possessing L. monocytogenes-specific IgG or after passive transfer of antibodies from primed pregnant, but not virgin, mice. Although maternal B cells were essential for producing IgGs that mediate vertically transferred protection, they were dispensable for antibody acquisition of protective function, which instead required sialic acid acetyl esterase5 to deacetylate terminal sialic acid residues on IgG variable-region N-linked glycans. Deacetylated L. monocytogenes-specific IgG protected neonates through the sialic acid receptor CD226,7, which suppressed IL-10 production by B cells leading to antibody-mediated protection. Consideration of the maternal-fetal dyad as a joined immunological unit reveals protective roles for antibodies against intracellular infection and fine-tuned adaptations to enhance host defence during pregnancy and early life.

Impaired malin expression and interaction with partner proteins in Lafora disease.

Lafora disease (LD) is an autosomal recessive myoclonus epilepsy with onset in the teenage years leading to death within a decade of onset. LD is characterized by the overaccumulation of hyperphosphorylated, poorly branched, insoluble, glycogen-like polymers called Lafora bodies. The disease is caused by mutations in either EPM2A, encoding laforin, a dual specificity phosphatase that dephosphorylates glycogen, or EMP2B, encoding malin, an E3-ubiquitin ligase. While glycogen is a widely accepted laforin substrate, substrates for malin have been difficult to identify partly due to the lack of malin antibodies able to detect malin in vivo. Here we describe a mouse model in which the malin gene is modified at the C-terminus to contain the c-myc tag sequence, making an expression of malin-myc readily detectable. Mass spectrometry analyses of immunoprecipitates using c-myc tag antibodies demonstrate that malin interacts with laforin and several glycogen-metabolizing enzymes. To investigate the role of laforin in these interactions we analyzed two additional mouse models: malin-myc/laforin knockout and malin-myc/LaforinCS, where laforin was either absent or the catalytic Cys was genomically mutated to Ser, respectively. The interaction of malin with partner proteins requires laforin but is not dependent on its catalytic activity or the presence of glycogen. Overall, the results demonstrate that laforin and malin form a complex in vivo, which stabilizes malin and enhances interaction with partner proteins to facilitate normal glycogen metabolism. They also provide insights into the development of LD and the rescue of the disease by the catalytically inactive phosphatase.

Histone H3 trimethylation at lysine 36 guides m6A RNA modification co-transcriptionally.

DNA and histone modifications have notable effects on gene expression1. Being the most prevalent internal modification in mRNA, the N6-methyladenosine (m6A) mRNA modification is as an important post-transcriptional mechanism of gene regulation2-4 and has crucial roles in various normal and pathological processes5-12. However, it is unclear how m6A is specifically and dynamically deposited in the transcriptome. Here we report that histone H3 trimethylation at Lys36 (H3K36me3), a marker for transcription elongation, guides m6A deposition globally. We show that m6A modifications are enriched in the vicinity of H3K36me3 peaks, and are reduced globally when cellular H3K36me3 is depleted. Mechanistically, H3K36me3 is recognized and bound directly by METTL14, a crucial component of the m6A methyltransferase complex (MTC), which in turn facilitates the binding of the m6A MTC to adjacent RNA polymerase II, thereby delivering the m6A MTC to actively transcribed nascent RNAs to deposit m6A co-transcriptionally. In mouse embryonic stem cells, phenocopying METTL14 knockdown, H3K36me3 depletion also markedly reduces m6A abundance transcriptome-wide and in pluripotency transcripts, resulting in increased cell stemness. Collectively, our studies reveal the important roles of H3K36me3 and METTL14 in determining specific and dynamic deposition of m6A in mRNA, and uncover another layer of gene expression regulation that involves crosstalk between histone modification and RNA methylation.

Developmental origin of the mammalian premaxilla.

The evolution of jaws has played a major role in the success of vertebrate expansion into a wide variety of ecological niches. A fundamental, yet unresolved, question in craniofacial biology is about the origin of the premaxilla, the most distal bone present in the upper jaw of all amniotes. Recent reports have suggested that the mammalian premaxilla is derived from embryonic maxillary prominences rather than the frontonasal ectomesenchyme as previously shown in studies of chicken embryos. However, whether mammalian embryonic frontonasal ectomesenchyme contributes to the premaxillary bone has not been investigated and a tool to trace the contributions of the frontonasal ectomesenchyme to facial structures in mammals is lacking. The expression of the Alx3 gene is activated highly specifically in the frontonasal ectomesenchyme, but not in the maxillary mesenchyme, from the beginning of facial morphogenesis in mice. Here, we report the generation and characterization of a novel Alx3CreERT2 knock-in mouse line that express tamoxifen-inducible Cre DNA recombinase from the Alx3 locus. Tamoxifen treatment of Alx3CreERT2/+;Rosa26mTmG/+ embryos at E7.5, E8.5, E9.5, and E10.5, each induced specific labeling of the embryonic medial nasal and lateral nasal mesenchyme but not the maxillary mesenchyme. Lineage tracing of Alx3CreERT2-labeled frontonasal mesenchyme from E9.5 to E16.5 clearly showed that the frontonasal mesenchyme cells give rise to the osteoblasts generating the premaxillary bone. Furthermore, we characterize a Dlx1-Cre BAC transgenic mouse line that expresses Cre activity in the embryonic maxillary but not the frontonasal mesenchyme and show that the Dlx1-Cre labeled embryonic maxillary mesenchyme cells contribute to the maxillary bone as well as the soft tissues lateral to both the premaxillary and maxillary bones but not to the premaxillary bone. These results clearly demonstrate the developmental origin of the premaxillary bone from embryonic frontonasal ectomesenchyme cells in mice and confirm the evolutionary homology of the premaxilla across amniotes.

Systemic administration of AAV-Slc25a46 mitigates mitochondrial neuropathy in Slc25a46-/- mice.

Mitochondrial disorders are the result of nuclear and mitochondrial DNA mutations that affect multiple organs, with the central and peripheral nervous system often affected. Currently, there is no cure for mitochondrial disorders. Currently, gene therapy offers a novel approach for treating monogenetic disorders, including nuclear genes associated with mitochondrial disorders. We utilized a mouse model carrying a knockout of the mitochondrial fusion-fission-related gene solute carrier family 25 member 46 (Slc25a46) and treated them with neurotrophic AAV-PHP.B vector carrying the mouse Slc25a46 coding sequence. Thereafter, we used immunofluorescence staining and western blot to test the transduction efficiency of this vector. Toluidine blue staining and electronic microscopy were utilized to assess the morphology of optic and sciatic nerves following treatment, and the morphology and respiratory chain activity of mitochondria within these tissues were determined as well. The adeno-associated virus (AAV) vector effectively transduced in the cerebrum, cerebellum, heart, liver and sciatic nerves. AAV-Slc25a46 treatment was able to rescue the premature death in the mutant mice (Slc25a46-/-). The treatment-improved electronic conductivity of the peripheral nerves increased mobility and restored mitochondrial complex activities. Most notably, mitochondrial morphology inside the tissues of both the central and peripheral nervous systems was normalized, and the neurodegeneration, chronic neuroinflammation and loss of Purkinje cell dendrites observed within the mutant mice were alleviated. Overall, our study shows that AAV-PHP.B's neurotrophic properties are plausible for treating conditions where the central nervous system is affected, such as many mitochondrial diseases, and that AAV-Slc25a46 could be a novel approach for treating SLC25A46-related mitochondrial disorders.

Mammalian germ cells are determined after PGC colonization of the nascent gonad.

Mammalian primordial germ cells (PGCs) are induced in the embryonic epiblast, before migrating to the nascent gonads. In fish, frogs, and birds, the germline segregates even earlier, through the action of maternally inherited germ plasm. Across vertebrates, migrating PGCs retain a broad developmental potential, regardless of whether they were induced or maternally segregated. In mammals, this potential is indicated by expression of pluripotency factors, and the ability to generate teratomas and pluripotent cell lines. How the germline loses this developmental potential remains unknown. Our genome-wide analyses of embryonic human and mouse germlines reveal a conserved transcriptional program, initiated in PGCs after gonadal colonization, that differentiates germ cells from their germline precursors and from somatic lineages. Through genetic studies in mice and pigs, we demonstrate that one such gonad-induced factor, the RNA-binding protein DAZL, is necessary in vivo to restrict the developmental potential of the germline; DAZL's absence prolongs expression of a Nanog pluripotency reporter, facilitates derivation of pluripotent cell lines, and causes spontaneous gonadal teratomas. Based on these observations in humans, mice, and pigs, we propose that germ cells are determined after gonadal colonization in mammals. We suggest that germ cell determination was induced late in embryogenesis-after organogenesis has begun-in the common ancestor of all vertebrates, as in modern mammals, where this transition is induced by somatic cells of the gonad. We suggest that failure of this process of germ cell determination likely accounts for the origin of human testis cancer.

Anthropoid primate-specific retroviral element THE1B controls expression of CRH in placenta and alters gestation length.

Pregnancy and parturition are intricately regulated to ensure successful reproductive outcomes. However, the factors that control gestational length in humans and other anthropoid primates remain poorly defined. Here, we show the endogenous retroviral long terminal repeat transposon-like human element 1B (THE1B) selectively controls placental expression of corticotropin-releasing hormone (CRH) that, in turn, influences gestational length and birth timing. Placental expression of CRH and subsequently prolonged gestational length were found in two independent strains of transgenic mice carrying a 180-kb human bacterial artificial chromosome (BAC) DNA that contained the full length of CRH and extended flanking regions, including THE1B. Restricted deletion of THE1B silenced placental CRH expression and normalized birth timing in these transgenic lines. Furthermore, we revealed an interaction at the 5' insertion site of THE1B with distal-less homeobox 3 (DLX3), a transcription factor expressed in placenta. Together, these findings suggest that retroviral insertion of THE1B into the anthropoid primate genome may have initiated expression of CRH in placental syncytiotrophoblasts via DLX3 and that this placental CRH is sufficient to alter the timing of birth.

Generation and characterization of Six2 conditional mice.

Heterozygous deletion of Six2, which encodes a member of sine oculis homeobox family transcription factors, has recently been associated with the frontonasal dysplasia syndrome FND4. Previous studies showed that Six2 is expressed in multiple tissues during craniofacial development in mice, including embryonic head mesoderm, postmigratory frontonasal neural crest cells, and epithelial and mesenchymal cells of the developing palate and nasal structures. Whereas Six2 -/- mice exhibited cranial base defects but did not recapitulate frontonasal phenotypes of FND4 patients, Six1 -/- Six2 -/- double mutant mice showed severe craniofacial defects including midline facial clefting. The complex phenotypes of FND4 patients and of Six1 -/- Six2 -/- mutant mice indicate that Six2 plays crucial roles in distinct cell types at multiple stages of craniofacial morphogenesis. Here we report generation of mice carrying insertions of a pair of loxP sites flanking exon-1 of the Six2 gene (Six2 f allele) using CRISPR/Cas9-mediated genome editing. We show that the Six2 f allele functions normally and is effectively inactivated by Cre-mediated recombination in vivo. Furthermore, we show that Six2 f/f ;Wnt1-Cre mice recapitulated cranial base defects but not neonatal lethality of Six2 -/- mice. These results indicate that Six2 f/f mice enable systematic investigation of cell type- and stage-specific Six2 function in development and disease.

The S52F FOXF1 Mutation Inhibits STAT3 Signaling and Causes Alveolar Capillary Dysplasia.

Rationale: Alveolar capillary dysplasia with misalignment of pulmonary veins (ACDMPV) is a lethal congenital disorder causing respiratory failure and pulmonary hypertension shortly after birth. There are no effective treatments for ACDMPV other than lung transplant, and new therapeutic approaches are urgently needed. Although ACDMPV is linked to mutations in the FOXF1 gene, molecular mechanisms through which FOXF1 mutations cause ACDMPV are unknown.Objectives: To identify molecular mechanisms by which S52F FOXF1 mutations cause ACDMPV.Methods: We generated a clinically relevant mouse model of ACDMPV by introducing the S52F FOXF1 mutation into the mouse Foxf1 gene locus using CRISPR/Cas9 technology. Immunohistochemistry, whole-lung imaging, and biochemical methods were used to examine vasculature in Foxf1WT/S52F lungs and identify molecular mechanisms regulated by FOXF1.Measurements and Main Results: FOXF1 mutations were identified in 28 subjects with ACDMPV. Foxf1WT/S52F knock-in mice recapitulated histopathologic findings in ACDMPV infants. The S52F FOXF1 mutation disrupted STAT3-FOXF1 protein-protein interactions and inhibited transcription of Stat3, a critical transcriptional regulator of angiogenesis. STAT3 signaling and endothelial proliferation were reduced in Foxf1WT/S52F mice and human ACDMPV lungs. S52F FOXF1 mutant protein did not bind chromatin and was transcriptionally inactive. Furthermore, we have developed a novel formulation of highly efficient nanoparticles and demonstrated that nanoparticle delivery of STAT3 cDNA into the neonatal circulation restored endothelial proliferation and stimulated lung angiogenesis in Foxf1WT/S52F mice.Conclusions: FOXF1 acts through STAT3 to stimulate neonatal lung angiogenesis. Nanoparticle delivery of STAT3 is a promising strategy to treat ACDMPV associated with decreased STAT3 signaling.

Knockout of latrophilin-3 in Sprague-Dawley rats causes hyperactivity, hyper-reactivity, under-response to amphetamine, and disrupted dopamine markers.

Attention deficit hyperactivity disorder is a pervasive developmental disorder characterized by inattention, impulsivity, and hyperactivity and is 75-90% heritable. Latrophilin-3 (LPHN3; or ADGRL(3)) is associated with a subtype of ADHD, but how it translates to symptoms is unknown. LPHN3 is a synaptic adhesion G protein coupled receptor that binds to fibronectin leucine rich transmembrane protein 3 and teneurin-3 (FLRT3 and TEN-3). We created a null mutation of Lphn3 (KO) in Sprague-Dawley rats using CRISPR/Cas9 to delete exon-3. The KO rats had no effects on reproduction or survival but reduced growth. KO females showed catch-up weight gain whereas KO males did not. We tested WT and KO littermates for home-cage activity, anxiety-like behavior, acoustic startle response, and activity after amphetamine challenge. Expression of Lphn3-related genes, monoamines, and receptors were determined. Lphn3 KO rats showed persistent hyperactivity, increased acoustic startle, reduced activity in response to amphetamine relative to baseline, and female-specific reduced anxiety-like behavior. Expression of Lphn1, Lphn2, and Flrt3 by qPCR and their protein products by western-blot analysis showed no compensatory upregulation. Striatal tyrosine hydroxylase, aromatic L-amino acid decarboxylase (AADC), and the dopamine transporter were increased and dopamine D1 receptor (DRD1) and dopamine- and cAMP-regulated neuronal phosphoprotein (DARPP-32) decreased with no changes in DRD2, DRD4, vesicular monoamine transporter-2, N-methyl-d-aspartate (NMDA)-NR1, -NR2A, or -NR2B. LPHN3 is expressed in many brain regions but its function is largely unknown. Data from human, mouse, zebrafish, Drosophila and our new Lphn3 KO rat data collectively show that its disruption is significantly correlated with hyperactivity and associated striatal changes in dopamine markers.

Loss of SLC25A46 causes neurodegeneration by affecting mitochondrial dynamics and energy production in mice.

Recently, we identified biallelic mutations of SLC25A46 in patients with multiple neuropathies. Functional studies revealed that SLC25A46 may play an important role in mitochondrial dynamics by mediating mitochondrial fission. However, the cellular basis and pathogenic mechanism of the SLC25A46-related neuropathies are not fully understood. Thus, we generated a Slc25a46 knock-out mouse model. Mice lacking SLC25A46 displayed severe ataxia, mainly caused by degeneration of Purkinje cells. Increased numbers of small, unmyelinated and degenerated optic nerves as well as loss of retinal ganglion cells indicated optic atrophy. Compound muscle action potentials in peripheral nerves showed peripheral neuropathy associated with degeneration and demyelination in axons. Mutant cerebellar neurons have large mitochondria, which exhibit abnormal distribution and transport. Biochemically mutant mice showed impaired electron transport chain activity and accumulated autophagy markers. Our results suggest that loss of SLC25A46 causes degeneration in neurons by affecting mitochondrial dynamics and energy production.

Licensing of primordial germ cells for gametogenesis depends on genital ridge signaling.

In mouse embryos at mid-gestation, primordial germ cells (PGCs) undergo licensing to become gametogenesis-competent cells (GCCs), gaining the capacity for meiotic initiation and sexual differentiation. GCCs then initiate either oogenesis or spermatogenesis in response to gonadal cues. Germ cell licensing has been considered to be a cell-autonomous and gonad-independent event, based on observations that some PGCs, having migrated not to the gonad but to the adrenal gland, nonetheless enter meiosis in a time frame parallel to ovarian germ cells -- and do so regardless of the sex of the embryo. Here we test the hypothesis that germ cell licensing is cell-autonomous by examining the fate of PGCs in Gata4 conditional mutant (Gata4 cKO) mouse embryos. Gata4, which is expressed only in somatic cells, is known to be required for genital ridge initiation. PGCs in Gata4 cKO mutants migrated to the area where the genital ridge, the precursor of the gonad, would ordinarily be formed. However, these germ cells did not undergo licensing and instead retained characteristics of PGCs. Our results indicate that licensing is not purely cell-autonomous but is induced by the somatic genital ridge.

A Transgenic Core Facility's Experience in Genome Editing Revolution.

The use of animal models, particularly rodents, has been immensely important to nearly all aspects of biomedical research, from basic science exploration to translational discoveries into clinical applications. The transgenic core facility that provides animal model production, preservation, and recovery services has been fundamental to the success of research efforts using animals. Recent advances in genome editing technologies, especially the clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated (Cas) enzyme system, have transformed the tedious animal model production into a simple and effective procedure. We, as a transgenic core facility established in 1993, adopted the CRISPR/Cas9 technology in early 2014 and have experienced the dramatic shift in the practice of animal model production, from the conventional embryonic stem cell approach to the direct genomic editing in rodent embryos. In this chapter, we describe the lessons that we learned from more than 200 genome editing projects performed in this core facility within the past 3 years. We also provide the practical guidelines for efficient generation of animal models using this technology and the insights into where new technologies lead us.

Retinoic acid activates two pathways required for meiosis in mice.

In all sexually reproducing organisms, cells of the germ line must transition from mitosis to meiosis. In mice, retinoic acid (RA), the extrinsic signal for meiotic initiation, activates transcription of Stra8, which is required for meiotic DNA replication and the subsequent processes of meiotic prophase. Here we report that RA also activates transcription of Rec8, which encodes a component of the cohesin complex that accumulates during meiotic S phase, and which is essential for chromosome synapsis and segregation. This RA induction of Rec8 occurs in parallel with the induction of Stra8, and independently of Stra8 function, and it is conserved between the sexes. Further, RA induction of Rec8, like that of Stra8, requires the germ-cell-intrinsic competence factor Dazl. Our findings strengthen the importance of RA and Dazl in the meiotic transition, provide important details about the Stra8 pathway, and open avenues to investigate early meiosis through analysis of Rec8 induction and function.

Tumor suppressor gene Rb is required for self-renewal of spermatogonial stem cells in mice.

The retinoblastoma tumor suppressor gene Rb is essential for maintaining the quiescence and for regulating the differentiation of somatic stem cells. Inactivation of Rb in somatic stem cells typically leads to their overexpansion, often followed by increased apoptosis, defective terminal differentiation, and tumor formation. However, Rb's roles in germ-line stem cells have not been explored. We conditionally disrupted the Rb gene in mouse germ cells in vivo and discovered unanticipated consequences for GFRa1-protein-expressing A(single) (GFRa1(+) A(s)) spermatogonia, the major source of male germ-line stem cells. Rb-deficient GFRa1(+) A(s) spermatogonia were present at normal density in testes 5 d after birth, but they lacked the capacity for self-renewal, resulting in germ cell depletion by 2 mo of age. Rb deficiency did not affect the proliferative activity of GFRa1(+) A(s) spermatogonia, but their progeny were exclusively transit-amplifying progenitor spermatogonia and did not include GFRa1(+) A(s) spermatogonia. In addition, Rb deficiency caused prolonged proliferation of progenitor spermatogonia, transiently enlarging this population. Despite these defects, Rb deficiency did not block terminal differentiation into functional sperm; offspring were readily obtained from young males whose germ cell pool was not yet depleted. We conclude that Rb is required for self-renewal of germ-line stem cells, but contrary to its critical roles in somatic stem cells, it is dispensable for their proliferative activity and terminal differentiation. Thus, this study identifies an unexpected function for Rb in maintaining the stem cell pool in the male germ line.

Gata4 is required for formation of the genital ridge in mice.

In mammals, both testis and ovary arise from a sexually undifferentiated precursor, the genital ridge, which first appears during mid-gestation as a thickening of the coelomic epithelium on the ventromedial surface of the mesonephros. At least four genes (Lhx9, Sf1, Wt1, and Emx2) have been demonstrated to be required for subsequent growth and maintenance of the genital ridge. However, no gene has been shown to be required for the initial thickening of the coelomic epithelium during genital ridge formation. We report that the transcription factor GATA4 is expressed in the coelomic epithelium of the genital ridge, progressing in an anterior-to-posterior (A-P) direction, immediately preceding an A-P wave of epithelial thickening. Mouse embryos conditionally deficient in Gata4 show no signs of gonadal initiation, as their coelomic epithelium remains a morphologically undifferentiated monolayer. The failure of genital ridge formation in Gata4-deficient embryos is corroborated by the absence of the early gonadal markers LHX9 and SF1. Our data indicate that GATA4 is required to initiate formation of the genital ridge in both XX and XY fetuses, prior to its previously reported role in testicular differentiation of the XY gonad.

Functional significance of the sex chromosomes during spermatogenesis.

Mammalian sex chromosomes arose from an ordinary pair of autosomes. Over hundreds of millions of years, they have evolved into highly divergent X and Y chromosomes and have become increasingly specialized for male reproduction. Both sex chromosomes have acquired and amplified testis-specific genes, suggestive of roles in spermatogenesis. To understand how the sex chromosome genes participate in the regulation of spermatogenesis, we review genes, including single-copy, multi-copy, and ampliconic genes, whose spermatogenic functions have been demonstrated in mouse genetic studies. Sex chromosomes are subject to chromosome-wide transcriptional silencing in meiotic and postmeiotic stages of spermatogenesis. We also discuss particular sex-linked genes that escape postmeiotic silencing and their evolutionary implications. The unique gene contents and genomic structures of the sex chromosomes reflect their strategies to express genes at various stages of spermatogenesis and reveal the driving forces that shape their evolution.Free Chinese abstract: A Chinese translation of this abstract is freely available at http://www.reproduction-online.org/content/149/6/R265/suppl/DC1.Free Japanese abstract: A Japanese translation of this abstract is freely available at http://www.reproduction-online.org/content/149/6/R265/suppl/DC2.